1163462951 Q * Wonka Ping timeout: 480 seconds 1163463518 Q * DavidS Ping timeout: 480 seconds 1163463856 Q * Aiken Quit: Leaving 1163464139 J * Aiken ~james@tooax8-089.dialup.optusnet.com.au 1163464418 M * Johnnie (This is a Gentoo question.) Does anyone happen to know when baselayout-vserver will be replaced with baselayout? 1163464431 M * Johnnie I've heard of plans and expectations, but I haven't seen anything yet. 1163466550 J * Akuma_Fan ~Akuma_Fan@58.247.6.149 1163469648 Q * fs hydrogen.oftc.net unununium.oftc.net 1163469749 J * fs fs@213.178.77.98 1163470107 Q * DreamerC_ Quit: leaving 1163470122 J * DreamerC ~dreamerc@61-224-132-244.dynamic.hinet.net 1163470533 N * Piet_ Piet 1163472622 Q * bronson Ping timeout: 480 seconds 1163474199 J * s0undt3ch_ ~s0undt3ch@81.193.56.192 1163474314 J * _dmax ~semaj@81.193.56.192 1163474657 Q * dmax Ping timeout: 480 seconds 1163474658 Q * s0undt3ch Ping timeout: 480 seconds 1163474660 N * _dmax dmax 1163474848 Q * s0undt3ch_ Ping timeout: 480 seconds 1163475211 J * s0undt3ch ~s0undt3ch@81.193.56.192 1163476908 M * doener Johnnie: AFAIK about 2 weeks ago 1163477047 M * doener Johnnie: http://planet.gentoo.org/developers/hollow/2006/11/03/baselayout_vserver_is_dead 1163477263 M * Loki|muh is it possible to enable vserver-scheduling on the fly? 1163477390 M * Loki|muh (i guess not, but it would be nice ;)) 1163477468 M * doener echo (hard|prio) > /proc/virtual/$XID/some-file 1163477472 M * doener or sth. like that 1163477488 M * doener might also be sched_hard/sched_prio 1163477575 M * doener or maybe via vattribute... had no coffee in the last few hours ;) 1163477603 M * Loki|muh hehe, no coffee here, too 1163477616 M * doener Loki|muh: yup, first use vsched then vattribute --xid --flag sched_hard 1163477616 M * Loki|muh tryin to sleep since 1am 1163477660 M * Loki|muh great! thanks! 1163477695 M * doener oh well, insert the actual xid between --xid and --flag *yawns* 1163477723 M * Loki|muh yeah, I worked that out after the first error message *g* 1163478069 Q * gerrit Quit: Client exiting 1163479482 Q * Piet Remote host closed the connection 1163479519 J * Piet hiddenserv@tor.noreply.org 1163480037 Q * ensc Killed (NickServ (GHOST command used by ensc_)) 1163480048 J * ensc ~irc-ensc@p54B4DA8F.dip.t-dialin.net 1163481908 J * gerrit ~gerrit@c-67-160-146-170.hsd1.or.comcast.net 1163482393 Q * Akuma_Fan Quit: Goodbye everyone. 1163483170 Q * Piet Quit: Piet 1163485596 J * meandtheshell ~markus@85-124-36-61.dynamic.xdsl-line.inode.at 1163485855 Q * Aiken Quit: Leaving 1163486832 J * samuel ~samuel@udp115186uds.hawaiiantel.net 1163486836 M * samuel hi 1163486877 J * bronson ~bronson@c-67-188-227-156.hsd1.ca.comcast.net 1163487056 M * Skram_ hello. 1163487333 M * samuel I'm looking for a part-time sysadmin (vserver, debian, postgres, lighttpd, ...). Most of the job can be done remotely, some on site (San Francisco (mostly), Fremont). Please, if you known someone, here is my contact info, samueltc@gmail.com 1163487484 M * Loki|muh if I would not live in Germany that would be interesting ;) 1163487735 M * samuel hehehe 1163490959 J * _cleric_ ~dacleric@p54823169.dip0.t-ipconnect.de 1163491171 J * dna_ ~naucki@146-243-dsl.kielnet.net 1163491314 M * renihs hehe 1163491317 M * renihs same :) 1163491368 Q * dacleric Ping timeout: 480 seconds 1163492641 J * Wonka produziert@chaos.in-kiel.de 1163492817 Q * daniel_hozac Ping timeout: 480 seconds 1163493262 Q * bronson Ping timeout: 480 seconds 1163494223 J * daniel_hozac ~daniel@c-2c1472d5.010-230-73746f22.cust.bredbandsbolaget.se 1163494782 J * DavidS ~david@vpn.uni-ak.ac.at 1163495177 J * Torsti76 ~irc@gate.iwm-kmrc.de 1163495202 M * Torsti76 Hi all! 1163495241 M * Torsti76 I have a strange problem on Gentoo 2006.1 with kernel 2.6.18 and vserver 2.1.1 1163495284 M * Torsti76 whenever i try to enter a guest with: 1163495284 M * Torsti76 vserver myvserver enter 1163495284 M * Torsti76 i get a 1163495284 M * Torsti76 vc_namespace: permission denied() 1163495308 M * Torsti76 the same with stop, start, restart, whatever 1163495334 M * Torsti76 funny thing is: at first, all vservers startup correctly via init-script 1163495350 M * Torsti76 which itself calls the vserver executable 1163495382 M * Torsti76 but as soon, as a new context is established, permission to it from the host system is completely denied 1163495397 M * Torsti76 does anyone know what's wrong here? 1163495565 M * DavidS you _are_ root? 1163495648 J * virtuoso ~s0t0na@shisha.spb.ru 1163496057 M * daniel_hozac Torsti76: lock in /etc/vservers/*/flags? 1163496528 J * meandtheshel1 ~markus@85-124-36-61.dynamic.xdsl-line.inode.at 1163496785 Q * meandtheshell Quit: Leaving. 1163496867 Q * meandtheshel1 Remote host closed the connection 1163496877 J * meandtheshell ~markus@85-124-36-61.dynamic.xdsl-line.inode.at 1163498194 M * Torsti76 yes, i'm root 1163498231 Q * eGnarF Read error: Connection reset by peer 1163498239 J * eGnarF ~bartek@bk.crystone.se 1163498324 M * doener Torsti76: what does cat /proc/self/status say? 1163498907 J * Aiken ~james@tooax6-137.dialup.optusnet.com.au 1163499054 M * Torsti76 i'll just have to see 1163499100 M * Torsti76 the problem is - the server is a production system, that i already downgraded to 2.0.2 again - there the problem does not occur 1163499115 M * daniel_hozac because 2.1.1 had a bug. 1163499128 M * daniel_hozac which is only exposed if you have lock in /etc/vservers/*/flags. 1163499136 M * daniel_hozac which doesn't really make sense anyway. 1163499142 M * Torsti76 ahh, that's the solution 1163499196 M * Torsti76 thanks, daniel - i'll try without lock, as soon as my second server (which is really for development) finished upgrading gcc, glibc, etc. 1163499213 M * doener daniel_hozac: bug as in "we finally supported it"? 1163499223 M * daniel_hozac no, bug as in we got the semantics slightly wrong ;) 1163499228 M * daniel_hozac http://vserver.13thfloor.at/Experimental/delta-private-fix01.diff fixes it. 1163499228 Q * Aiken Read error: Connection reset by peer 1163499265 M * Torsti76 daniel, that's great - so i can send a bugfixed ebuild to gentoo's bugzilla 1163499627 M * daniel_hozac or just ask Hollow or phreak`` to put it in there ;) 1163499723 M * Torsti76 ok, hope they read it... 1163499734 M * Torsti76 i already updated my local ebuild 1163499749 M * Torsti76 let's see what happens 1163500167 Q * Borg- Ping timeout: 480 seconds 1163500189 J * Borg- borg@cube.benet.uu3.net 1163500727 Q * samuel Ping timeout: 480 seconds 1163501659 J * Aiken ~james@tooax6-137.dialup.optusnet.com.au 1163501954 Q * Borg- Quit: Lost terminal 1163502719 M * matti :) 1163503678 J * lilalinux ~plasma@dslb-084-058-193-119.pools.arcor-ip.net 1163504449 Q * Aiken Quit: Leaving 1163505809 M * Torsti76 the patch causes the whole host to hang! 1163505829 M * Torsti76 do i have to remove the lock flag in any case? 1163505831 M * daniel_hozac which patch? 1163505836 M * daniel_hozac delta-private-fix01? 1163505987 J * eyck_ qv6PJDiS@nat.nowanet.pl 1163506098 Q * eyck Ping timeout: 480 seconds 1163506750 M * Torsti76 yes, with the delta-private-fix01 1163506780 M * Torsti76 i applied it, afterwards the vserver starts just hung during boot 1163506793 M * Torsti76 i.e. timed out 1163506865 M * Torsti76 then i removed the flags file in each vs-config (they only contained "lock" ;o), which resulted in an kernel oops saying vc_net_create: permission denied() 1163506873 M * Torsti76 so, i just reverted to 2.0.2 1163506905 M * Torsti76 because i already have a bunch of angry users with torches and heyforks behind my back... ;o) 1163508875 Q * mnemoc Ping timeout: 480 seconds 1163509306 Q * shedi Quit: Leaving 1163509329 J * mnemoc ~amery@kilo105.server4you.de 1163510010 Q * mire_ Remote host closed the connection 1163510630 M * Loki|muh someone knows if there is a need for sun sparcstations for vserver-testing purposes? 1163510970 J * Piet hiddenserv@tor.noreply.org 1163511098 Q * mnemoc Ping timeout: 480 seconds 1163511524 J * mnemoc ~amery@kilo105.server4you.de 1163512077 J * maharaja maharaja@is.the.one.who.rules.at 1163512082 M * maharaja hi there 1163512095 M * maharaja long time no see 1163512443 M * daniel_hozac hello 1163513037 M * maharaja ive got a question regarding vservers and lvm 1163513044 M * maharaja someone allready mentioned it on the mailinglist 1163513048 M * daniel_hozac oh? 1163513204 M * maharaja 1sec 1163513214 M * maharaja # lvchange -an /dev/vg0/testlv 1163513214 M * maharaja LV vg0/ipax_2 in use: not removing 1163513221 M * maharaja ups 1163513225 M * maharaja # lvchange -an /dev/vg0/ipax_2 1163513225 M * maharaja LV vg0/ipax_2 in use: not removing 1163513235 M * maharaja mixed the wrong lines ;) 1163513242 M * maharaja vserver2 2.6.16.19-vs2.0.2-rc21 1163513313 M * daniel_hozac and you've read the solutions of unmounting it in all the namespaces? 1163513362 J * marcfiu ~mef@aegis.CS.Princeton.EDU 1163513368 M * marcfiu good morning 1163513390 M * Johnnie Any Gentoo geeks in here? 1163513400 M * Johnnie I have a Gentoo specific question. 1163513412 M * maharaja yes 1163513433 M * Johnnie Alright, I asked about baselayout last night sometime, and doener replied with the following link...: 1163513434 M * maharaja i even did a "for i in `seq 1 200`; do chcontext --xid $i unount /dev/vg0/ipax_2; done" 1163513444 M * Johnnie http://planet.gentoo.org/developers/hollow/2006/11/03/baselayout_vserver_is_dead 1163513461 M * Johnnie And I was wondering if anyone knew the status of baselayout 1.13. 1163513523 M * Johnnie maharaja: I've had the same problem, and I've had to unmount namespaces as daniel_hozac suggests. 1163513543 M * Johnnie I've had better luck with dlimits, rather than sizing partitions. 1163513544 M * Hollow Johnnie: 1.13.0_alpha5 1163513550 J * Borg- borg@cube.benet.uu3.net 1163513556 M * Johnnie You have more of a dynamic method of handling space. 1163513564 M * Johnnie Oh okay, Hollow ... excellent. 1163513591 M * Johnnie I was curious, I had erratic Asterisk issues with newer versions... 1163513598 M * maharaja Johnnie: what exactly did you do? 1163513663 M * Johnnie Check this out, maharaja... http://oldwiki.linux-vserver.org/Namespaces 1163513702 M * Johnnie I 1163513703 M * Johnnie err 1163513714 M * Johnnie I've found that sometimes you have to unmount all of the namespaces. :-/ 1163513721 M * daniel_hozac that's expected. 1163513737 M * daniel_hozac you have to unmount it from every namespace created since you mounted it. 1163513765 M * Johnnie Okay, I guess that makes sense. 1163513772 M * Johnnie I never put a lot of thought to it. 1163513774 M * Johnnie haha 1163513777 M * daniel_hozac that's why you're supposed to put guest specific mounts in the guest specific fstab ;) 1163513786 M * Johnnie Gotcha. 1163513794 M * Johnnie That's a handy feature. 1163513814 M * Johnnie Before I discovered that was available, I had each guest using their own portage tree...yikes. 1163513814 M * daniel_hozac 0.30.211 has some basic namespace cleanup features present though. 1163513853 M * daniel_hozac meaning, if you enable it, mounts it won't be able to access or might need access to, will be unmounted automaticall. 1163513856 M * daniel_hozac +y 1163513857 M * maharaja ah, ok i forgot that namespaces != context 1163513870 M * Johnnie Right, maharaja. 1163513876 M * Johnnie daniel_hozac: Excellent. 1163513878 M * maharaja and thought that using "chcontext" did set the namespace" 1163513891 M * maharaja how do i get an overview of all namespaces? 1163513900 M * daniel_hozac you don't, really. 1163513905 M * maharaja is the namespce the xid? 1163513927 M * daniel_hozac sort of. 1163513934 M * daniel_hozac you can associate a namespace with a context. 1163513971 M * maharaja how do i know which namespaces exist? 1163514082 M * maharaja and does a vnamespace -e XXX umount /dev/vg0/ipax_2 suffice? 1163514127 M * maharaja done that: for i in `seq 1 20000`; do vnamespace -e $i umount/dev/vg0/ipax_2 |grep -i ipax_2; done 1163514130 M * maharaja does not help 1163514407 M * maharaja i always get "vnamespace: vc_enter_namespace(): No such process" which i assume is the correct behaviour when the namespace does not exist 1163514521 M * maharaja ah, i missed a space 1163514595 M * maharaja anyways, this does not resolve my issue 1163514690 M * maharaja might there be something with the devicemapper 1163514828 M * daniel_hozac unlikely. 1163514874 M * daniel_hozac for i in `ls -1 /proc/virtual`; do [ -d $i ] || continue; vnamespace -e $i umount /dev/vg0/ipax_2; done 1163514877 M * daniel_hozac or something like that. 1163515303 M * maharaja its not working :-/ 1163515706 M * maharaja # file /dev/vg0/unidev 1163515706 M * maharaja /dev/vg0/unidev: symbolic link to `/dev/mapper/vg0-unidev' 1163515710 M * maharaja (i renamed it) 1163515896 N * Bertl_zZ Bertl 1163515901 M * Bertl morning folks! 1163516060 M * marcfiu hey bertl 1163516079 M * Bertl ah, saw your 'request' they just got moved 1163516114 M * marcfiu the scheduler diffs... ok. 1163516144 M * Bertl i.e. I moved them to some subdirectory 1163516155 M * marcfiu ok 1163516218 M * Bertl i.e. check out Experimental/OLD-DELTA/ 1163516296 M * maharaja daniel_hozac: any idea whats wrong? 1163516301 M * marcfiu got it 1163516302 M * marcfiu thanks 1163516369 M * Bertl maharaja: what did you rename? 1163516525 M * Bertl daniel_hozac: still around? 1163516648 J * Piet_ hiddenserv@tor.noreply.org 1163516932 M * maharaja Bertl: i have troubles unmounting a lvm volume 1163516941 M * maharaja i did try to unmount it in all xids 1163516948 M * maharaja i did try to unmount it in all vnamespaces 1163516982 M * maharaja then i try a "lvchange -an /dev/vg0/ipax_2" and it state that the device is busy and cannot be set inactive 1163516986 M * Torsti76 Bertl, hi! 1163517054 M * Torsti76 Bertl: Did you read my msgs about kernel-oops with gentoo and 2.1.1 (even with delta-private-fix01)? 1163517067 M * Bertl maharaja: then clearly somebody is using it, or you have a kernel bug 1163517094 Q * Piet Ping timeout: 480 seconds 1163517098 M * Bertl Torsti76: could be, did you upload the oops? 1163517131 M * Torsti76 nope, it occured during restart - i had no possibilities to get it from the console... 1163517150 M * Bertl that's what serial console is for :) 1163517163 M * Torsti76 and i can't freely experiment, since it's a production server 1163517177 M * Torsti76 and my folks are already killin' me ;o) 1163517205 M * Bertl Torsti76: okay, so what is it you expect us to do? :) 1163517236 M * Torsti76 just wanted to know, whether it is a known problem 1163517255 M * Bertl at least with mainline, everything works smooth here 1163517278 M * Torsti76 and i have a second server - that's for development, but it needs some "surrounding" updates, before i can test 1163517280 M * Bertl as I don 1163517295 M * Bertl +'t use gentoo kernels, better ask Hollow 1163517314 M * Torsti76 ok, i'll get back to him... 1163517320 M * Bertl but IIRC, the latest tree should be in sync 1163517336 M * Torsti76 ... as soon as i finished upgrading the 2nd server 1163517363 M * Bertl if you can reproduce the issue and/or provide a stack dump, we can investigate 1163517369 M * Torsti76 with 2.0.2, everything works perfekt 1163517378 M * Torsti76 ok, i'll try my best! 1163517437 M * Bertl maharaja: did you configure something ontop of the lvm? 1163517473 M * Torsti76 but it can't be some general kernel option, like "compat vdso support"? (just to be sure ;o) 1163517480 M * Bertl maharaja: maybe a vroot device or so? 1163517520 M * Bertl Torsti76: there are many many options, my kernel have compat vdso enabled too .. without more information, it's hard to tell 1163517562 M * Torsti76 ok, i'll come up with a proper crash dump ;o) 1163517580 M * Bertl that'd be excellent! 1163517599 M * Bertl if it is a general issue, maybe you can trigger it on a laptop or so 1163517639 Q * Piet_ Ping timeout: 480 seconds 1163517662 M * Torsti76 oh, the 2nd server will be just fine for that (and a dual xeon compiles a bit faster than a centrino ;o) 1163517760 M * maharaja Bertl: no, i did not configure anything on top. /dev/vg0/ipax_2 is a link to /dev/mapper/vg0-ipax_2 1163517784 M * maharaja below lvm (pv0) is a raid1 md device 1163517815 M * maharaja kernel is: 2.6.16.19-vs2.0.2-rc21, tools are: 0.30.208 1163517839 M * maharaja the ipax_2 vserver ist of course stopped and in the host system the lv is unmounted 1163517864 M * Bertl okay, and did you try the suggested command sequence (from daniel)? 1163517974 M * maharaja yes, the for; ... ; done did not initially work (dont know why) but i ran the commands by hand 1163518034 M * maharaja there is something strange though: 1163518057 M * maharaja i renamed ipax_2 to unidev and created /vservers/unidev to mount the new /dev/vg0/unidev there 1163518065 M * maharaja # rmdir ipax_2 1163518065 M * maharaja rmdir: `ipax_2': Device or resource busy 1163518139 M * Bertl that sounds like there is something mounted there 1163518149 M * Bertl check what /proc/mounts has to say 1163518205 M * maharaja mhm 1163518241 M * maharaja mhm... 1163518253 M * maharaja umount /dev/vg0/ipax_2 did not work, but umount /vservers/ipax_2 did 1163518267 M * maharaja (in every namespace) 1163518269 M * maharaja strange thing 1163518288 M * Bertl hmm, strange umount command 1163518296 M * Bertl probably it checks /etc/mtab or so 1163518884 Q * node Quit: Lost terminal 1163519392 M * maharaja anyways, if this is handled by the utils then they should take that into account 1163519452 M * Bertl recent utils 0.30.211 should handle that properly 1163519469 M * Bertl 0.30.208 is more than two years old? 1163519564 M * maharaja yes, thats possible :) 1163519575 M * maharaja production environment ... i try to upgrade as seldom as possible 1163519609 M * Bertl that's fine, but don't expect that we fix bugs/add features by time travel :) 1163519625 M * Bertl although it would be an interesting approach :) 1163520139 M * Bertl daniel_hozac: okay, I uploaded a version (for 2.6.19-rc5) which fixes the proc stuff (not cleaned up yet) and is supposed to handle the namespaces (especially uts) properly 1163520165 M * Bertl daniel_hozac: you either need to clone() with the new flags or apply the delta-clone-hack01.diff too 1163520174 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.19-rc5-vs2.1.1-t2.diff 1163520191 M * Bertl I'm off now, translocating .. should be back later 1163520195 N * Bertl Bertl_oO 1163520256 M * maharaja Bertl_oO: i just wanted to point that out for future releases :) 1163520361 J * shedi ~siggi@cpe-24-165-167-167.midsouth.res.rr.com 1163521022 M * daniel_hozac Bertl_oO: ok, cool! 1163521176 M * daniel_hozac maharaja: as i said, namespace cleanup should've already solved it. 1163521335 M * maharaja daniel_hozac: yes, too bad that we did not get that unmount did not work as expected 1163521361 M * daniel_hozac i thought you got it to work? 1163521435 M * maharaja yes, by umount /path/to/mountpoint 1163521440 M * maharaja not by /dev/ 1163521466 M * daniel_hozac right. 1163521526 M * daniel_hozac my command was a bit off anyway, it needed /proc/virtual in [ -d $i ]. 1163521584 J * stefani ~stefani@tsipoor.banerian.org 1163522014 M * maharaja daniel_hozac: i fixed that, dont worry 1163522040 M * maharaja umount /dev/vg0/ipax_2 did not work, umount /vserves/ipax_2 did 1163522142 M * daniel_hozac right. 1163522895 M * k3mper how should i disable mounting /tmp as 16MB tmpfs? 1163522944 M * k3mper ops /etc/mtab :) 1163522968 M * k3mper hmm it's still there 1163523069 M * m4z k3mper: is it not in /etc/vservers//fstab ? 1163523088 M * k3mper no 1163523113 M * m4z (in the host server) 1163523206 M * m4z like this: 1163523207 M * m4z cat /etc/vservers/apache/fstab 1163523207 M * m4z none /proc proc defaults 0 0 1163523207 M * m4z none /tmp tmpfs size=16m,mode=1777 0 0 1163523208 M * m4z none /dev/pts devpts gid=5,mode=620 0 0 1163523215 M * m4z (sorry for spam) 1163523256 M * k3mper sorry it's there. (am blind :) 1163523260 M * k3mper thanks 1163523733 M * m4z hehe, no problem 1163524269 J * bonbons ~bonbons@83.222.36.93 1163524622 M * Hollow http://linux-vserver.org/VServer_Hosting rocks. it just got me an internship :) 1163524680 M * daniel_hozac Hollow: nice! 1163524683 M * Rich_Estill Hollow: Congrats. Internships rock. 1163524689 M * daniel_hozac i should start looking for one myself... 1163524694 M * Hollow i hope it's paid :P 1163524717 M * Rich_Estill I wish my boss would let me have an intern. I just had to go and get my own pizza. 1163524725 M * Hollow heh 1163524749 M * daniel_hozac Bertl_oO: so this'll go into vnamespace, correct? 1163524849 M * daniel_hozac Hollow: so where did you get it? 1163524859 M * Hollow MPeX Networks 1163524893 M * daniel_hozac cool. 1163524987 A * phreak`` looks at Hollow 1163525000 M * Hollow phreak``: lazy bitch! :P 1163525017 M * phreak`` Hollow: yeah, still owe you a disk, I know :P 1163525036 M * Hollow heh 1163525039 M * Hollow gtg, cu later 1163525045 M * daniel_hozac cya. 1163525049 M * phreak`` Hollow: have fun 1163525087 M * daniel_hozac phreak``: did you see Torsti76's report earlier today? why do Gentoo users have lock in flags? :) 1163525126 M * matti Hollow, phreak``, daniel_hozac :) 1163525132 M * daniel_hozac hello matti. 1163525189 M * phreak`` daniel_hozac: no I didn't see that :) (have been pretty busy with a public tender for SAN-storage/SAN-network); but wasn't that what I encountered ? :) 1163525202 M * daniel_hozac yep. 1163525215 M * daniel_hozac so i'm wondering why Gentoo users have that ;) 1163525252 M * daniel_hozac or is it just a coincidence? 1163525294 M * phreak`` daniel_hozac: as I told you when I encountered that (and you asked me exactly the same - where did I get that from); it was in some old Gentoo Howto (iirc) 1163525319 M * daniel_hozac hmm, sorry, i don't remember that. 1163525328 M * phreak`` daniel_hozac: no need to be sorry ;) 1163525329 M * daniel_hozac but i guess that explainsi t. 1163525359 M * daniel_hozac i thought such legacy flags were limited to Debian though :) 1163525367 M * phreak`` heh :) 1163525424 M * daniel_hozac is a fixed version in the overlay already? 1163525463 M * phreak`` daniel_hozac: of what ? :) the howto ? its even on our project site :) http://www.gentoo.org/proj/en/vps/vserver-howto.xml 1163525477 M * daniel_hozac of the patch ;) 1163525491 M * phreak`` which patch ? *me looks a bit confused* 1163525513 M * daniel_hozac 2.1.1. 1163525541 M * daniel_hozac http://vserver.13thfloor.at/Experimental/delta-private-fix01.diff is at least supposed to fix it. 1163525581 M * phreak`` guess not, I updated the svn this morning and doesn't look like Hollow added that :) 1163525683 M * phreak`` daniel_hozac: anything else for 2.1.1 ? 1163525716 M * daniel_hozac http://vserver.13thfloor.at/Experimental/delta-vlimit-clean01.diff i suppose, but that's not as severe. 1163525749 M * daniel_hozac (and i'm not sure it's been tested too much yet) 1163525877 M * phreak`` daniel_hozac: so better not add it for the time being ? 1163525901 M * daniel_hozac well, IMHO it should be safe. 1163525937 M * daniel_hozac i'll give it a spin though to make sure. 1163526079 M * phreak`` daniel_hozac: much appreciated :) 1163526848 J * MrX ~urk@219.95.237.112 1163527667 Q * DavidS Ping timeout: 480 seconds 1163527983 M * daniel_hozac i must be doing something wrong. 1163528004 M * daniel_hozac 2.1.1+delta-private-fix01+delta-vlimit-fix01 gives me a BUG() on chbind. 1163529362 Q * gerrit Ping timeout: 480 seconds 1163529837 M * Hollow phreak``, daniel_hozac: well, the lock flag is not related to gentoo IMO, maybe some other howto contains it, but it's not in our howtos/scripts/patches or whatever... regarding the fix, it is not in yet, i usually don't add deltas to our patchset, i just wait for the next version... 1163529951 M * phreak`` well I ocasionally do add delta's (once they are "safe" and "tested") 1163530233 M * Hollow daniel_hozac: btw, will util-vserver use the new uts namespaces in 2.6.19? (just reading the backlog) 1163530715 M * daniel_hozac i hope so. 1163531069 M * daniel_hozac Hollow: do you remember what you wanted to get cleaned up in vlimit? 1163531480 J * yarihm ~yarihm@84-75-123-221.dclient.hispeed.ch 1163531700 J * gerrit ~gerrit@bi01p1.co.us.ibm.com 1163533143 M * daniel_hozac ok, delta-private-fix01 is broken. 1163533699 J * DavidS ~david@diode.black.co.at 1163533926 M * yarihm hey guys ... 1163533985 M * Hollow daniel_hozac: support for all limits in 2.1.1? :) 1163534008 M * daniel_hozac that's already there :) 1163534029 M * daniel_hozac hello yarihm. 1163534087 M * yarihm i've a question regarding limits of the memory a vserver can allocate. is there something like a guaranteed minimum and a maximum that can be set? i've checked the docs on the wiki but somehow ... well, i didn't get it i guess 1163534110 M * daniel_hozac you can't do guarantees. 1163534166 M * yarihm daniel_hozac: other than having a server with more ram than the maximum set for all vservers summed up i guess 1163534179 M * yarihm (provided that the host itself isn't going havoc) 1163534180 M * daniel_hozac right. 1163534276 M * yarihm thanks, i suspected so 1163534316 M * yarihm (implementing a minimum works for things like network-traffic, but what would you do if someone wanted his RAM back while you are using it? kill your stuff or what?) :) 1163534343 J * samuel ~samuel@udp115186uds.hawaiiantel.net 1163534374 M * daniel_hozac Hollow: only thing missing in util-vserver's 2.1 support are the missing syscalls. 1163534383 M * daniel_hozac (AFAIK, please correct me if i'm wrong) 1163534462 M * DavidS yarihm: well, one yould always swap out someone else as long as sum(min_limits) < phys. RAM 1163534504 M * DavidS or rather ... not swap out if min watermark reached 1163534760 J * comfrey ~comfrey@adsl-ull-84-47.44-151.net24.it 1163534954 M * yarihm DavidS: yeah, that would be a nice feature ... 1163535052 M * DavidS yarihm: you could tune /proc/sys/vm/swappiness, this helps at least globally ... 1163535060 Q * Rich_Estill Quit: Leaving 1163535402 M * daniel_hozac Bertl_oO, Hollow, phreak``: http://people.linux-vserver.org/~dhozac/p/k/delta-private-fix02.diff 1163535503 M * daniel_hozac and the vlimit cleanup seems to work fine. 1163536039 M * phreak`` daniel_hozac: hrm is that incremental to the previous, or a replacement ? :) 1163536052 M * phreak`` ah, incremental :P 1163536178 M * Hollow daniel_hozac: didn't look too closely lately, but i saw some 2.1.1 changes, so i guess it went in :) 1163536236 M * Hollow looks good to me as is :) 1163536286 M * daniel_hozac phreak``: there's still a bug in vc_net_create though. 1163536306 M * phreak`` daniel_hozac: hrm ? 1163536313 M * Hollow and we still need the syscall for accounted limits :) 1163536353 M * daniel_hozac if nx_migrate_task fails (i.e. with -EACCES), vc_net_create will trigger a BUG_ON in __unhash_vx_info. 1163536384 M * daniel_hozac (it's not holding the lock it requires) 1163536411 M * daniel_hozac __unhash_nx_info, obviously. 1163536468 Q * comfrey Ping timeout: 480 seconds 1163536528 M * daniel_hozac i'm not sure if we should just take the lock in vc_net_create, or use another function for it. 1163536645 J * pzYsTorM schak@dslc-082-082-079-063.pools.arcor-ip.net 1163536723 M * pzYsTorM hey folks. ever seen a bug like this? http://nopaste.biz/?9590 1163536832 M * daniel_hozac export LC_ALL=C LANG=C 1163536839 M * daniel_hozac ;) 1163536852 M * pzYsTorM hehe ;p 1163536933 M * pzYsTorM shall i translate it for you on the fly? ;p 1163536954 M * Hollow EBUSY 1163536957 M * Hollow ;) 1163537020 M * daniel_hozac http://linux-vserver.org/Namespaces 1163537059 M * daniel_hozac every vserver started since you mounted it will have a copy of the mount. 1163537081 M * Hollow we should make a big warning in there "don't mount devices as guest root if you painlessly want to unmount it lateron" ;) 1163537119 M * daniel_hozac or maybe enable namespace cleanup by default ;) 1163537134 M * Hollow is it safe? 1163537146 M * daniel_hozac what do you mean by safe? 1163537172 M * daniel_hozac what's in trunk shouldn't unmount too many things it needs. 1163537180 A * Hollow goes looking ;) 1163537197 M * daniel_hozac plus it lets you add paths to the list of skipped directories. 1163537208 M * pzYsTorM i was searching a method of implementing a hard quota for a guest. and container were at that time the only chance. 1163537236 M * daniel_hozac hmm, disk limits have been around for years. 1163537250 J * boodie ~boodie@host245-73-dynamic.8-87-r.retail.telecomitalia.it 1163537312 M * pzYsTorM hmm... and the per-user-quota inside a guest? 1163537329 P * boodie 1163537369 M * daniel_hozac for shared devices? not implemented on 2.6 due to lack of interest. 1163537430 M * pzYsTorM hm ok. these work with the loopback-mount of the container pretty good. 1163537491 M * pzYsTorM could you explain what you meant with: "every vserver started since you mounted it will have a copy of the mount." ... why a copy of the mount? 1163537531 M * daniel_hozac because namespaces are cloned. 1163537537 M * daniel_hozac and with it, all mounts. 1163537561 M * pzYsTorM although they didnt have the device to access it? 1163537568 M * Hollow daniel_hozac: yeah, looks neat :) 1163537596 M * Hollow took some time to figure the logic .. 1163537605 M * daniel_hozac hehe, indeed. 1163537609 M * daniel_hozac thus the comments ;) 1163537654 M * Hollow bash is horrible :P 1163537666 M * daniel_hozac could've been perl :) 1163537678 M * Hollow heh 1163537712 M * Hollow well, haskell is not better 1163537769 M * Wonka haskell rocks 1163537779 M * Hollow yeah, but the syntax sucks :D 1163537828 A * phreak`` stabs Hollow with some Haskell 1163537830 M * Hollow well, i just started to learn haskell, but somehow i don't like its syntax ;) 1163537844 M * Hollow but list comprehension is neat 1163537861 M * phreak`` Hollow: :P 1163537891 M * Hollow phreak``: pah! 1163537935 M * Hollow while $disk_not_arrived; sleep 1; done 1163537953 M * Hollow i forgot a 'do' 1163537958 M * Hollow :) 1163537959 M * Hollow sorry bash 1163538088 Q * eyck_ Ping timeout: 480 seconds 1163538104 M * phreak`` Hollow: lol, I'll hopefully get to check the drives (my spare ones) tomorrow, or get a trainee to do that :P after that I'll send you one (or maybe two) 1163538137 M * Hollow heh, yeah.. as long as my box boots everything is still ok :P 1163538399 J * eyck kJuyqoXc@nat.nowanet.pl 1163538635 J * Aiken ~james@tooax8-162.dialup.optusnet.com.au 1163542134 Q * virtuoso Ping timeout: 480 seconds 1163542222 M * gdm hi, a little problem that i'm probably being stupid about 1163542237 M * gdm i created a vserver on an ip, didn't use it and then stopped it 1163542255 M * gdm then, i removed the /etc/vservers/VSERVER directory and contents 1163542282 M * gdm and when i tried to remove the /var/lib/vservers/VSERVER directory, it said it was busy 1163542303 M * gdm and now, i have a new vserver on that ip, but when starting it i get the following line: 1163542318 M * gdm RTNETLINK answers: File exists 1163542326 M * gdm how can i fix this? 1163542344 M * daniel_hozac "didn't use it"? 1163542356 M * daniel_hozac sounds like stopping it failed though. 1163542394 M * gdm but now i have removed the config directory, so is there anyway i can just kill whatever is still running? 1163542410 M * gdm or, even, figure out what is still running? 1163542447 M * daniel_hozac vps faux doesn't list anything? 1163542466 M * gdm nope, not under the old name 1163542534 M * daniel_hozac obviously, but check the xids. 1163542540 M * daniel_hozac you did use a static xid, didn't you? :) 1163542588 Q * bonbons Quit: Leaving 1163542594 M * gdm yeah, i did... hmm, just gotta remember it ;-) 1163542717 M * gdm no, nothing 1163542801 M * daniel_hozac so no directory in /proc/virtual either? did you install the guest on a separate mount? 1163542830 M * gdm yes, it was a separate mount 1163542836 Q * meandtheshell Quit: Leaving. 1163542844 M * daniel_hozac and you didn't put it in the guest's fstab? 1163542868 M * daniel_hozac (man, 3 people in the same day... namespace cleanup is sooo going to be enabled by default in 0.30.212) 1163542910 J * mire ~mire@168-167-222-85.adsl.verat.net 1163542953 M * gdm daniel_hozac: er, didn't put what in the gues's fstab? sorry for being a dumbass... :/ 1163542957 Q * DavidS Ping timeout: 480 seconds 1163542961 M * daniel_hozac that mount. 1163542970 M * daniel_hozac the guest's /-mount. 1163543028 M * gdm no... i didn't. i didn't realise i had to... i guess i better go do some reading 1163543041 M * gdm cos i've done this for a lot of vservers now - all the same error, i guess 1163543109 M * daniel_hozac yes, because all mounts are copied into the guest's namespace when you start it. 1163543120 M * daniel_hozac so all of your running guests have references to that mount. 1163543170 M * daniel_hozac you'll have to unmount it in all of them before you can remove it. 1163543186 M * gdm er - i have been creating a new partition for each guest 1163543241 M * daniel_hozac per-guest quotas? 1163543378 M * gdm yeah, that's what i'm trying to get 1163543386 M * gdm to limit disk space per guest 1163543422 J * dna___ ~naucki@17-208-dsl.kielnet.net 1163543460 M * daniel_hozac you don't need per-user/group quotas inside? 1163543498 P * stefani I'm Parting (the water) 1163543509 M * daniel_hozac because you can easily do per-guest limits (disk limits) and still get the benefits of hashification/unification. 1163543513 M * gdm no, at least not right now. it maybe the case that in the fuure 1163543543 M * gdm daniel_hozac: ah, really? i was under the impression that to do per gues limits, they needed to be on unique partitions 1163543552 M * gdm obviously i was wrong 1163543584 M * daniel_hozac that's only for per-user/group quota that you want to control from the inside. 1163543695 M * gdm ok. so how do i do it otherwise? 1163543710 M * gdm i guess i haven't seen the right page on teh wiki 1163543750 Q * Greek0 Read error: Connection reset by peer 1163543765 J * Greek0 ~greek0@85.255.145.201 1163543770 M * daniel_hozac http://oldwiki.linux-vserver.org/Disk+Limits 1163543853 Q * dna_ Ping timeout: 480 seconds 1163543855 M * gdm ok. thx. i think i didn't understand that before. that all refers to fstab ont eh host, doesn't it 1163544027 M * daniel_hozac yes. 1163544241 M * pzYsTorM it is not advised to mount the rootfs with tagxid, but having one loopback-mount for 3 guests is better than having 3 loopback-mounts for 3 guests? 1163544297 M * daniel_hozac how many mounts you need depends on your requirements, as i said. 1163544312 M * daniel_hozac if you need quota controllable from the inside, you need one mount/device per guest. 1163544321 M * daniel_hozac (quota == per-user/group) 1163544370 M * daniel_hozac but having separate mounts comes with drawbacks, e.g. you can't use hashification/unification to make your guests use less space (disk and RAM). 1163544473 M * pzYsTorM ah okay. it sounds so buggy and vulnarable to use one mount/user... when you told me 2 hrs ago. 1163544662 M * daniel_hozac there are some aspects of it that people seem to forget, such as the namespaces. 1163544690 M * daniel_hozac so if you're going to use one mount per guest, i strongly suggest putting it in the guest's fstab, or using namespace cleanup. 1163544704 Q * dna___ Quit: Verlassend 1163544720 M * pzYsTorM yes of course. thats the intension, that a guest can check his quota with df 1163544752 M * pzYsTorM how could a guest check the quota when using the dlimit method` 1163544753 M * pzYsTorM ? 1163544782 M * daniel_hozac df, of course. 1163544821 M * pzYsTorM but how... i cant make a fstab entry... and so df says nothing 1163544883 M * daniel_hozac because you always have a root mount. 1163544886 M * pzYsTorM thinking of a not-existint device? 1163544890 M * pzYsTorM -i+e 1163544906 M * daniel_hozac df asks the kernel about the usage, and the kernel returns the virtualized values. 1163545154 M * pzYsTorM 1 year ago i tested this. and df told me nothing. but ok. perhaps i did something wrong. 1163545233 M * pzYsTorM this "namespace cleanup" - how do i do this? is there a command? vcontext --help says nothing about cleanup 1163545256 M * pzYsTorM args lol 1163545267 M * daniel_hozac http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1163545284 M * pzYsTorM ok to late vnamespace --cleanup found 1163545292 M * daniel_hozac that's not it. 1163545297 M * pzYsTorM oh? 1163545302 M * daniel_hozac support for that has been removed in the kernel. 1163545325 M * daniel_hozac and in the most recent versions where it was still included, it just oopsed your kernel :) 1163545336 M * pzYsTorM lol. ok 1163545371 M * pzYsTorM i love this great flower page. but it doesnt tell me, what should write into these files? only create them with no contents? 1163545396 M * pzYsTorM especially "namespace-cleanup" 1163545464 M * daniel_hozac it's a bool, so just having it exist is sufficient. 1163545503 Q * mnemoc Ping timeout: 480 seconds 1163545630 Q * Aiken Remote host closed the connection 1163545632 J * Aiken ~james@tooax8-162.dialup.optusnet.com.au 1163545990 J * mnemoc ~amery@kilo105.server4you.de